Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 40648
    Prev
    Next
    INDEX
    Explanations

    IL

    np_max-act · gemini-2.0-flash

    This neuron activates on short token fragments that form part of biochemical or molecular names (e.g. drug names, cytokines, proteins).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     teklif
    -0.07
     toes
    -0.07
    학과
    -0.06
     eBay
    -0.06
    Compatibility
    -0.06
    .PRO
    -0.06
    bonus
    -0.06
     Lonely
    -0.06
     lg
    -0.06
    КИ
    -0.06
    POSITIVE LOGITS
    -react
    0.07
    Doctrine
    0.07
    categoria
    0.06
    CJK
    0.06
    velocity
    0.06
    Þ
    0.06
     اب
    0.06
     contenders
    0.05
     osób
    0.05
    [pos
    0.05
    Activations Density 0.077%

    No Known Activations